Case StudyUS Food and Drug Administration (FDA)

US Food and Drug Administration (FDA) automates conversion of manually submitted forms to digital format with 91% accuracy using deep neural networks

The FDA receives 10% of applications in a hand-written format. After employing seasonal contractors to enter these manually into digital format, they have now implemented Captricity''s AI powered READ engine to automate conversion of manual forms to digital format. The program is 1000s of times faster and matches human accuracy.

Context

"The FDA receives hundreds of thousands of applications every year out of which 10% is estimated to be manual."

The Project

One of the databases developed backlogs every year as 10% of the submissions were manual. FDA approached Captricity to automate capturing the information contained in the manual documents.

Data

"Trained on millions of real-world examples using language and visual machine learning models"

Results

Automation reduces manual staff requirements by 75% with and accuracy of 90-99%.

Related Use Cases

Use Case

Recognise unstructured text, e.g. handwriting, in documents to extract information to streamline processes like account creation, loan and insurance origination and documentation

→

←Back to Case Studies